Long-term effectiveness of recommended boosted protease inhibitor-based antiretroviral therapy in Europe.

Methods: Data were analysed for 5678 EuroSIDA-enrolled patients starting a DRV/r-, ATZ/r- or LPV/r-containing regimen between 1 January 2000 and 30 June 2013. Separate analyses were performed for the following subgroups of patients: (1) ART-naïve subjects (8%) at ritonavir-boosted protease inhibitor (PI/r) initiation; (2) ART-experienced individuals (44%) initiating the new PI/r with a viral load (VL) ≤500 HIV-1 RNA copies/mL; and (3) ART-experienced patients (48%) initiating the new PI/r with a VL >500 copies/mL. Virological failure (VF) was defined as two consecutive VL measurements >200 copies/mL ≥24 weeks after PI/r initiation. Kaplan-Meier and multivariable Cox models were used to compare risks of failure by PI/r-based regimen. The main analysis was performed with intention-to-treat (ITT) ignoring treatment switches.

Results: The time to VF favoured DRV/r over ATZ/r, and both were superior to LPV/r (log-rank test; P 0.02) in all analyses. Nevertheless, the risk of VF in ART-naïve patients was similar regardless of the PI/r initiated after controlling for potential confounders. The risk of VF in both treatment-experienced groups was lower for DRV/r than for ATZ/r, which, in turn, was lower than for LPV/r-based ART.

Conclusion: The time to VF favoured DRV/r over ATZ/r, and both were superior to LPV/r (log-rank test; P 0.02) in all analyses. Nevertheless, the risk of VF in ART-naïve patients was similar regardless of the PI/r initiated after controlling for potential confounders. The risk of VF in both treatment-experienced groups was lower for DRV/r than for ATZ/r, which, in turn, was lower than for LPV/r-based ART.